As with all by-elections, transferability of the result to general elections remains very circumstantial.
Those who infer universal truths from who is elected to Parliament from a handful of seats are often failing to take into account local flavour, riding issues or the higher impact the candidates themselves play.
They may also be missing the changed dynamic because no one is selecting a Prime Minister, and no one is selecting a government program.
Finally, and most importantly, the lower media coverage and public awareness massively dampens both enthusiasm and turnout.
But the results of the recent by-elections still have meaning. They tell us something very fundamental about our four major political parties.
First of all, let's talk about exceptions. Some by-elections are actually referenda on the government, which can have a predictive factor for the next elections.
Deb Gray's success in Beaver River in 1989 presaged that Reform had the ability to win seats, and posed a serious threat to the PC Party.
Others are ballots on a specific local issue.
For instance, the Ancaster-Dundas-Flamborough-Aldershot by-election in Ontario was a referendum on Hamilton amalgamation. The governing PCs lost and a very safe PC seat flipped massively Liberal.
The infamous Hastings by-election of 1934 was a referendum on funding Catholic schools, and was a resounding no. (This episode of sectarian bigotry in Ontario's not-so-distant past is a key element in understanding George Drew's failure as national Tory leader in the 1950s, the defeat of Frank Miller and the 42-year Ontario PC dynasty in 1985, and just what happened to John Tory in 2007.)
But those "referendum by-elections" are the exception, not the rule.
In fact, many by-elections that are seen as "referendum by-elections" may not have actually been fought that way.
The best example is the oft-cited Newport by-election of 1922 in Britain.
Basically, during the First World War, David Lloyd George had split the governing Liberal Party and led a faction into coalition government with the support of the Tories.
At the end of the war, Lloyd George was immensely popular for "winning the war" and the Conservatives continued to subsume themselves to him for fear that his popularity was the only thing keeping them in office. Almost by accident, the Conservatives ran a candidate against the Lloyd George Liberals in this by-election, which the Conservatives unexpectedly won.
The result was the Conservatives believing they were electable on their own, withdrawing from the coalition and eventually becoming the dominant party in Britain throughout most of the 20th century.
As a result of its huge historic impact, the Newport by-election was ret-conned into a referendum on continuing the Coalition. But the reality is the Newport by-election was fought on local issues like temperance and who was the most local man. All three candidates opposed continuing the coalition, and the issue never really came up.
It is only in the context of what happened next that the Newport by-election takes on its "referendum" status.
Those are the exceptions. The truth of by-elections is actually very mundane.
By-elections tend to be about whose vote decays the least.
Basically, turnout drops by about half during a by-election, sometimes a little less and sometimes a lot more.
The question that decides by-elections more often than not is which party was able to turn out the most people who supported them last time. Sometimes turnout can be dramatically suppressed but the suppression is even.
During a snow storm in Scarborough Rouge-River, a safe Liberal seat looked to be in jeopardy because no one - no one - was voting. Turnout was 18 per cent, but the Liberals still won. Vote decay was basically even for all the candidates because the cause was external (weather).
But this is not a test of external factors, but the ability of parties to overcome lower media attention and voter interest to get those predisposed to their party to actually trot out to the polls and vote.
The recent four by-elections show us that two parties have figured out how to do this, and two have not.
The Conservatives are excellent at keeping their voters engaged with politics between elections.
Conservative-leaning voters and activists are bombarded with fundraising letters, ten-percenters, e-mails and advertising.
In Montmagny-L'Islet-Kamouraska-Rivière-du-Loup, the Conservatives were basically able to get the same 13,000 people who voted for them in 2008 to come back in 2009. The Bloc could only get half of the 20,000 people who voted for them last time to the polls. That's the election right there.
(This is not always the case, as the Conservatives did an atrocious job of getting their vote out in Hochelaga.)
But another party has joined the Conservatives in being able to mobilize their vote between elections.
In B.C., the NDP was able to get their vote to the polls in New Westminster-Coquitlam.
This was a marked difference from by-elections in 2008, which saw the NDP vote share fall precipitously, especially in the two Toronto races.
Clearly, the NDP has worked on its outreach and intra-election mobilization. As Brian Topp points out, the NDP turned out two-thirds of their general election vote in these by-elections, which is no mean feat.
But there are two other parties who simply aren't matching the CPC and NDP in intra-election organization.
The Bloc Quebecois humiliated itself in this election, with more than 23,000 votes walking out the door in just two seats. In contrast, even the Liberals only lost 11,000 votes of a much larger base.
This is evidence that the BQ is really hurting organizationally in Quebec.
The culprit could be the growing divide in Quebec between a techno-bureaucratic, baby-boomer elite that is hugely supportive of the welfare state as the source of their wealth and power, and younger generations that are increasingly skeptical of the PQ and BQ. With an increasingly grey-bearded core vote, it's tough to find the youthful legs to run people to the polls.
The Liberals are also falling dangerously behind.
While none of these were ridings where the Liberals expect to compete in the near future, winning isn't everything in politics.
The subsidy for political parties is based on total votes, and so votes in uncompetitive ridings are almost as important as those in tight seats. A 308-seat strategy isn't just about making the grassroots feel good; it's about building a successful 21st century political party around the contemporary rules of the game.
More ominously, the Liberals failed to finish second in any of the four. In 2000, they finished second in every one.
At issue is the decline of the Liberals as a truly national party. Where once the Grits were in play in just about every place in the country, today they are projecting to finish first or second in fewer than half the seats.
Say what you will about the Liberal Party of Canada, but it is a truly national institution. Its decline from that position threatens the ability of political parties to internally broker the conflicts in Canadian politics, and instead compete as vote blocs in a fractured Parliament. That is very different Canada from the one we have lived in for the past 150 years, where either the Liberals or PCs acted as brokerage parties, reducing regional conflict.
There are lessons that come from these by-elections, and they matter.